Re: Very uneven distribution of packages per maintainer

2003-05-27 Thread Colin Watson 
On Tue, May 27, 2003 at 10:40:06PM +0200, Yann Dirson wrote:
> Adam wrote:
> > So doing bts work is worthless?  :)
> 
> Hey, someone once wrote similar scripts to count how many bugreports
> were reported by anyone !

Try:

  http://bugs.debian.org/cgi-bin/pkgindex.cgi?indexon=submitter&sortby=count

(Beware, the output is over a megabyte.)

> /me rejoices recalling he was ranked 3rd by the number of open bugs :)

You still are. :)

Cheers,

-- 
Colin Watson  [EMAIL PROTECTED]

Re: Very uneven distribution of packages per maintainer

2003-05-27 Thread Yann Dirson 
Adam wrote:
> So doing bts work is worthless?  :)

Hey, someone once wrote similar scripts to count how many bugreports
were reported by anyone !

/me rejoices recalling he was ranked 3rd by the number of open bugs :)

Well, never mind :)

-- 
Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ?
Debian-related: <[EMAIL PROTECTED]> |   Support Debian GNU/Linux:
Pro:<[EMAIL PROTECTED]> |  Freedom, Power, Stability, Gratuity
 http://ydirson.free.fr/| Check

Re: Very uneven distribution of packages per maintainer

2003-05-26 Thread Michael Banck 
On Sat, May 24, 2003 at 09:15:39AM -0400, Ben Collins wrote:
> Even that is little indicator. I know folks who's job has nothing to do
> with Debian, but they do more work than people who are paid to work on
> Debian. It's all about motivation, and no person is doing anything
> wrong. It's a personal choice how much time and energy you devote to
> Debian. As long as that number isn't zero, then you are welcome to stay.

Amen.

Michael

Re: Very uneven distribution of packages per maintainer

2003-05-25 Thread Jamie Wilkinson 
This one time, at band camp, Petter Reinholdtsen wrote:
>Packages  Developers

Looks like a normal distribution curve in gnuplot.

-- 
[EMAIL PROTECTED]   http://people.debian.org/~jaq

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Adam Heath 
On Sat, 24 May 2003, Wouter Verhelst wrote:

> On Fri, May 23, 2003 at 09:48:02PM -0500, Adam Heath wrote:
> > On Sat, 24 May 2003, Wouter Verhelst wrote:
> >
> > > The script can't even get everything a Debian Developer does for Debian.
> > > While most, if not all, active Debian Developers do packaging work,
> > > there's other stuff to be done -- such as taking care of autobuilders,
> > > being a sysadmin, ftp-master, listadmin, or release manager, doing
> > > porters' work.
> >
> > So doing bts work is worthless?  :)
>
> Mind the 'such as'. These are examples; I'm sure I forgot hundreds of
> other people doing valuable non-packaging work.
>
> Should've been more explicit, perhaps.

That was as joke.

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Paul Seelig 
[EMAIL PROTECTED] (Petter Reinholdtsen) writes:

> [Tollef Fog Heen]
> > So, why do you think having a more even distribution is a good
> > thing?
> 
> Because in Debian there is a few people with high "load" in debian,
> and many with less "load".
>
I think this is the wrong way to see it.  Since work for Debian is
based on a voluntary basis developers who have mor packages to process
can allow themselves to invest more time and effort into such an
undertaking than others. I'd rather propose to think the other way
round: People with a higher "load" are more likely to work on less
packages and people with a lesser "load" have more time/energy/money
left to work on more packages.
 
> Of course, the packages per developer is not a perfect indicator of
> "load", but it is an indicator.
> 
Indicating that their maintainers can afford investing more of his/her
time, effort, money, etc. than other people? Yes, definitely!

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread John Hasler 
Petter Reinholdtsen wrote:
> Because in Debian there is a few people with high "load" in debian,
> and many with less "load".  People with high load are more likely to
> burn out and disappear.

Do you have statistics to support that statement?
-- 
John Hasler
[EMAIL PROTECTED] (John Hasler)
Dancing Horse Hill
Elmwood, WI

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Ben Collins 
On Sat, May 24, 2003 at 11:25:03PM +1000, Russell Coker wrote:
> On Sat, 24 May 2003 22:15, Petter Reinholdtsen wrote:
> > Because in Debian there is a few people with high "load" in debian,
> > and many with less "load".  People with high load are more likely to
> > burn out and disappear.  It is thus better to have more people with
> > less load.
> >
> > Of course, the packages per developer is not a perfect indicator of
> > "load", but it is an indicator.
> 
> Sometimes you can have an easy source package that produces several other 
> packages.  Sometimes you can have a difficult source package that only 
> produces one package.
> 
> Then there's the issue of upstream maintainers who are also Debian 
> developers, 
> being the active upstream developer can be enough work that developing any 
> other Debian packages would take too much time.
> 
> Even if you had a metric for measuring the amount of work a developer was 
> doing, that wouldn't gain you anything.  We can't force people who have a 
> small number of packages to take on more work, and we don't want to 
> discourage them from contributing altogether.

Not to mention those that are paid to work on Debian (Hello HP :) and
those who have full time jobs that are irrelevant to anything they do
for Debian. Then we could get into single/married, kids, hobbies, etc.

Even that is little indicator. I know folks who's job has nothing to do
with Debian, but they do more work than people who are paid to work on
Debian. It's all about motivation, and no person is doing anything
wrong. It's a personal choice how much time and energy you devote to
Debian. As long as that number isn't zero, then you are welcome to stay.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
Deqo   - http://www.deqo.com/

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Russell Coker 
On Sat, 24 May 2003 22:15, Petter Reinholdtsen wrote:
> Because in Debian there is a few people with high "load" in debian,
> and many with less "load".  People with high load are more likely to
> burn out and disappear.  It is thus better to have more people with
> less load.
>
> Of course, the packages per developer is not a perfect indicator of
> "load", but it is an indicator.

Sometimes you can have an easy source package that produces several other 
packages.  Sometimes you can have a difficult source package that only 
produces one package.

Then there's the issue of upstream maintainers who are also Debian developers, 
being the active upstream developer can be enough work that developing any 
other Debian packages would take too much time.

Even if you had a metric for measuring the amount of work a developer was 
doing, that wouldn't gain you anything.  We can't force people who have a 
small number of packages to take on more work, and we don't want to 
discourage them from contributing altogether.

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Wouter Verhelst 
On Fri, May 23, 2003 at 09:48:02PM -0500, Adam Heath wrote:
> On Sat, 24 May 2003, Wouter Verhelst wrote:
> 
> > The script can't even get everything a Debian Developer does for Debian.
> > While most, if not all, active Debian Developers do packaging work,
> > there's other stuff to be done -- such as taking care of autobuilders,
> > being a sysadmin, ftp-master, listadmin, or release manager, doing
> > porters' work.
> 
> So doing bts work is worthless?  :)

Mind the 'such as'. These are examples; I'm sure I forgot hundreds of
other people doing valuable non-packaging work.

Should've been more explicit, perhaps.

-- 
Wouter Verhelst
Debian GNU/Linux -- http://www.debian.org
Nederlandstalige Linux-documentatie -- http://nl.linux.org
"An expert can usually spot the difference between a fake charge and a
full one, but there are plenty of dead experts." 
  -- National Geographic Channel, in a documentary about large African beasts.


pgpRT7sgYPuQb.pgp
Description: PGP signature

Re: Very uneven distribution of packages per maintainer

2003-05-24 Thread Petter Reinholdtsen 
[Tollef Fog Heen]
> So, why do you think having a more even distribution is a good
> thing?

Because in Debian there is a few people with high "load" in debian,
and many with less "load".  People with high load are more likely to
burn out and disappear.  It is thus better to have more people with
less load.

Of course, the packages per developer is not a perfect indicator of
"load", but it is an indicator.

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Adam Heath 
On Fri, 23 May 2003, Ben Collins wrote:

> "Oh no, the Collins family has 2 acres per child...those poor kids! Uh,
> you say that 3.5 acres of your yard is a lake?"

Dredging.  With a snorkel.

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Adam Heath 
On Sat, 24 May 2003, Wouter Verhelst wrote:

> The script can't even get everything a Debian Developer does for Debian.
> While most, if not all, active Debian Developers do packaging work,
> there's other stuff to be done -- such as taking care of autobuilders,
> being a sysadmin, ftp-master, listadmin, or release manager, doing
> porters' work.

So doing bts work is worthless?  :)

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Wouter Verhelst 
On Fri, May 23, 2003 at 05:26:28PM -0700, Sean 'Shaleh' Perry wrote:
> >
> > Not necessarily -- some packages are a lot of work, like xfree, glibc,
> > apache, some are a decent amount of work, like mailman, cvs and some
> > are close to zero work, like chrpath and xslide.  People also have
> > different amounts of time available -- those who are paid to do Debian
> > maintainence at work will have more time than somebody who works 12
> > hours/day without any Debian work in there, etc.
> >
> > So, why do you think having a more even distribution is a good thing?
> > Or rather, why is the current situation so bad?
> >
> 
> indeed.  Some packages are "worth" 10 "normal" packages in the amount of work 
> they require.
> 
> Also, perhaps the script could deal with items like qa owning orphaned 
> packages and the like.

The script can't even get everything a Debian Developer does for Debian.
While most, if not all, active Debian Developers do packaging work,
there's other stuff to be done -- such as taking care of autobuilders,
being a sysadmin, ftp-master, listadmin, or release manager, doing
porters' work.

The work that's being done which does not involve "being listed as a
package's maintainer" is not to be underestimated...

-- 
Wouter Verhelst
Debian GNU/Linux -- http://www.debian.org
Nederlandstalige Linux-documentatie -- http://nl.linux.org
"An expert can usually spot the difference between a fake charge and a
full one, but there are plenty of dead experts." 
  -- National Geographic Channel, in a documentary about large African beasts.


pgpbbtO9j2tWn.pgp
Description: PGP signature

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Ben Collins 
On Fri, May 23, 2003 at 11:53:05PM +0200, Petter Reinholdtsen wrote:
> 
> I just ran some stats on my APT sources (mostly Woody), and discovered
> that the distribution of number of packages per developer is very
> uneven.  This is the histogram of developers with the specific number
> of packages they maintain:

This is terrible! Wait, I forgot, these numbers mean absolutely nothing.

I'm not sure what you think these numbers represent, but it's like
taking a ratio for each family in a neighborhood of acres/child and
deciding if those kids have too much yard work.

"Oh no, the Collins family has 2 acres per child...those poor kids! Uh,
you say that 3.5 acres of your yard is a lake?"

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
Deqo   - http://www.deqo.com/

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Matt Zimmerman 
On Fri, May 23, 2003 at 05:26:28PM -0700, Sean 'Shaleh' Perry wrote:

> > So, why do you think having a more even distribution is a good thing?
> > Or rather, why is the current situation so bad?
> 
> indeed.  Some packages are "worth" 10 "normal" packages in the amount of work 
> they require.

And a maintainer who only has time for one package should maintain one
package, and do a decent job of it, rather than overextend himself.

-- 
 - mdz

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Sean 'Shaleh' Perry 
>
> Not necessarily -- some packages are a lot of work, like xfree, glibc,
> apache, some are a decent amount of work, like mailman, cvs and some
> are close to zero work, like chrpath and xslide.  People also have
> different amounts of time available -- those who are paid to do Debian
> maintainence at work will have more time than somebody who works 12
> hours/day without any Debian work in there, etc.
>
> So, why do you think having a more even distribution is a good thing?
> Or rather, why is the current situation so bad?
>

indeed.  Some packages are "worth" 10 "normal" packages in the amount of work 
they require.

Also, perhaps the script could deal with items like qa owning orphaned 
packages and the like.

Re: Very uneven distribution of packages per maintainer

2003-05-23 Thread Tollef Fog Heen 
* Petter Reinholdtsen 

| I suspect a more even distribution of packages per maintainer would be
| a good thing. :)

I think your stats are wrong, at least they are different from
http://www.debian.gr.jp/~kitame/maint.cgi

Not necessarily -- some packages are a lot of work, like xfree, glibc,
apache, some are a decent amount of work, like mailman, cvs and some
are close to zero work, like chrpath and xslide.  People also have
different amounts of time available -- those who are paid to do Debian
maintainence at work will have more time than somebody who works 12
hours/day without any Debian work in there, etc.

So, why do you think having a more even distribution is a good thing?
Or rather, why is the current situation so bad?

-- 
Tollef Fog Heen,''`.
UNIX is user friendly, it's just picky about who its friends are  : :' :
  `. `' 
`-

Very uneven distribution of packages per maintainer

2003-05-23 Thread Petter Reinholdtsen 

I just ran some stats on my APT sources (mostly Woody), and discovered
that the distribution of number of packages per developer is very
uneven.  This is the histogram of developers with the specific number
of packages they maintain:


Packages  Developers
   1 239
   2 128
   3 112
   4  89
   5  72
   6  55
   7  33
   8  20
   9  34
   10 19
   11 16
   12 16
   13 16
   14  8
   15 10
   16  9
   17  7
   18  7
   19  2
   20  6
   21  5
   22  5
   23  3
   24  3
   25  3
   26  3
   27  3
   29  3
   32  1
   33  1
   34  1
   37  1
   40  1
   41  1
   43  2
   44  1
   46  1
   47  1
   50  2
   52  1
   65  2
   76  1
   83  1
   84  1

One developer is listed as the maintainer of 84 packages.  I am
impressed. :)

24 developers are responsible for 1004 packages in Debian.

I suspect a more even distribution of packages per maintainer would be
a good thing. :)


This is the script I used to generate the histogram.

#!/usr/bin/perl -w

sub lowercase_email {
$maint = shift;
($name, $addr) = $maint =~ m/^(.+)<(.+)>$/;
#return "$name <\L$addr>";
return "$name";
}

open(APT, "apt-cache dumpavail |") || die "Unable to run apt-cache";
while () {
chomp;
$pkg = $1 if /^Package: (.+)$/;
$maint = $1 if /^Maintainer: (.+)$/;
$src = $1 if /^Source: (.+)$/;
if (/^$/) {
$pkg = $src if ($src);
#print "P: $pkg M: $maint\n";
if ( ! exists $pkgs{$pkg}) {
$pkgs{$pkg} = 1;
$maint = lowercase_email($maint);
$maintainers{$maint}++;
}

undef $src;
}
}
close(APT);

for $maint (keys %maintainers) {
#print "$maintainers{$maint} $maint\n";
$hist{$maintainers{$maint}}++;
}

for $count (sort { $a <=> $b } keys %hist) {
print "$count $hist{$count}\n";
}